A resale side hustle's first-month take-home: ¥19,591. A DIY craftsman's full 5-month record of ¥730,000 in sales, ¥170,000 in profit

Laying out five months of actual figures as-is

Someone who moved to a mountainous area of Nagano and runs a business as a “DIY proxy” started resale (“sedori”) as a side hustle in June 2025. What’s notable is that they published monthly sales, cost, profit, and margin on note every single month, continuing for five months — and then wrote a post announcing they were pausing. Plenty of people document how they started. Few leave a record, complete with numbers, of the slowdown and the decision to step back.

Average monthly sales were ¥145,968, average monthly profit ¥34,108. The target the author originally set was “¥50,000 profit per month”, a mark none of the five months reached.

For that same June, incidentally, the paid (¥150) “Month One” report shows different figures: sales of ¥55,733, cost of ¥36,142, net profit of ¥19,591, margin 35.2%. There’s a ¥1,074 gap in profit versus the free monthly report. It appears to stem from timing of the closing date or when fees were reflected, but it also illustrates how much “profit for the month” can swing in a personal side-business’s bookkeeping.

What they sourced, and what actually made money

The category is secondhand physical goods, power tools, cameras, sporting goods. The most profitable item in month one was a Makita air nailer (AF502N), sourced for ¥3,300 and sold for ¥13,000, a profit of ¥8,400 at a 64.6% margin. And this was purchased as “unverified junk,” then function-tested by the author personally before listing.

July’s top item was an OLYMPUS XA2: sourced at ¥18,700, sold at ¥42,000, profit ¥19,030. In September, a SONY α57 double-zoom kit brought ¥7,394 in profit. In October, a Hi-KOKI chainsaw (CS3625DC) brought ¥7,981, and a MAX pin nailer (TA-245) brought ¥4,350 at a 62.1% margin. The winning pattern was consistent: “buy tools and cameras cheap that might be broken, verify them, and sell them.”

The time invested in sourcing is also disclosed. Month one’s in-store sourcing amounted to effectively one day’s worth (two half-days). They stopped by a shop in Komoro on the way back from accompanying their spouse, who had thrown out their back, to a medical appointment, and separately made a round of shops around Nagano city on another day. Learning came not from books or courses but purely from YouTube and ChatGPT. The author writes that their process became: “when a category or item catches my interest, I consult ChatGPT first before researching.”

The tide turned in August, the month they switched to “online sourcing only”

This case has a clear turning point: August. With their day job busy, and amid the summer heat, they couldn’t find time to make the rounds of physical shops, so they limited sourcing to online only. The result showed up strongly in the numbers.

Sales rose from July’s ¥116,352 to ¥267,389, a 2.3x jump. But profit actually dipped slightly, from ¥47,957 to ¥43,521. Margin fell from 41.2% to 16.3%, less than half. Sourcing cost ballooned from ¥68,395 to ¥223,868, a 3.3x increase, meaning nearly all the added sales were consumed by sourcing cost.

The author’s own analysis of the loss is specific: “items that looked pristine in photos turned out to be junk in person.” Many items labeled “function unverified” turned out to be genuinely broken, producing multiple loss-making sales.

What had actually been working was “the ability to inspect”

Why did in-store sourcing yield 41% while online sourcing dropped to 16%? On the surface it’s tempting to explain it as “online sourcing is more competitive,” but for this person, a simpler structure was at work.

What they had actually been earning was less a price gap than the discount margin attached to uncertainty about an item’s condition. Tools listed as “function unverified” or “junk” get discounted heavily by sellers to offload risk. Someone whose day job is DIY proxy work can pick up such a tool and judge, by hand, whether it’s actually broken, whether it can be fixed, and how many minutes that fix would take. The ¥3,300 AF502N becoming ¥13,000 is the market paying for exactly that judgment.

Online sourcing strips away the very venue where that skill can be exercised. Once the only judgment material is a listing photo and text, the edge from being a good judge of condition disappears, and all that’s left is the price-gap search anyone can do. That’s why sales could still grow while margin converged toward an industry-average level. Rather than “the sourcing method changed”, the truer description is “their own weapon got disarmed.” August’s 16.3% is that number, told plainly.

A different kind of distortion appeared in September. In order to clear unsold inventory, the author switched to prioritizing quick sales, which increased the share of low-margin items and pulled margin down to 14.0%. Sales held at ¥200,800, but profit was ¥28,180. Conversely, October narrowed sourcing and listings, and margin recovered to 36.0%, but sales dropped to ¥89,901, with profit landing at ¥32,364. Sourcing that builds sales and sourcing that leaves profit turn out to be different things, and one person’s working hours can’t run both at once, which is the reality these five months demonstrate.

The reason it didn’t continue wasn’t “it doesn’t pay”

In December, the monthly reports paused. The reasons the author cited: average profit had settled around ¥20,000-30,000 and become repetitive. The day job was busy and left no time for sourcing, and the snowboard season left no free time. And, bluntly: “the numbers barely change anymore, and it’s not fun.” This wasn’t a loss-driven withdrawal so much as a case of no longer finding a reason to spend discretionary time on work with no visible room to grow.

The resale activity itself hasn’t stopped. In an April 2026 post, the author discloses a breakdown from that winter’s sourcing: of 6 action cameras, 4 sold at a 24% margin, hiking gear at 46%, tools at 55%. A newly explored category, imported figurines via Yahoo Auctions, at 42%. And a UL backpack bought cheap due to a pricing error at 82%. At the same time, they also posted losses of -40% on tools and -30% on damaged goods, noting that repair costs can exceed the item’s original cost, and expressing anxiety about an unsold camera worth over ¥100,000. It settled into a position of: “not making money, but there’s learning here that differs from my day job.”

How much of this can be replicated

What’s replicable is the recordkeeping habit. Because sales, cost, profit, and margin were reported in the same format every single month, the anomaly of “sales up 2.3x, profit flat” in August became visible both to the author and to readers. Had it been run on gut feel alone, it likely would have registered simply as “a great month where sales nearly quintupled.” Keeping learning to just YouTube and ChatGPT, and keeping initial investment near zero, are also replicable.

What’s harder to replicate is the core skill of judging condition. Their day job as a DIY proxy, and their knowledge of power-tool construction and repair effort, was the source of the margin. Someone without that domain knowledge using the same method would find August’s 16.3% becoming their normal. Living in a rural area where you can drive around recycle shops, and having a lifestyle that allows half-day sourcing trips tacked onto errands or hospital visits, are also part of the conditions.

And the biggest limit is scale. Against a stated target of ¥50,000 profit per month, the average was ¥34,108. The five-month total profit of ¥170,539 is thin compensation for the labor of inspecting tools, taking photos, packing, and shipping around a day job. What gives this case value is that this thinness wasn’t hidden. It was recorded in numbers.

Related reading

• A case of launching a kitchen car in 4 months — a record tracking the initial cost and ramp-up of a small business handling physical goods
• An engineer’s side-business rental space: ¥7.5 million in year one across 3 locations — the income statement of a side business that chose not to sell its time by the hour